On Fri, May 10, 2019 at 01:45:47PM +0000, Chris Mason wrote: > I'm a little concerned about about btrfs_csum_data() and > btrfs_csum_final() below. We're using two different > SHASH_DESC_ON_STACK() and then overwriting internals (shash_desc_ctx()) > with the assumption that whatever we're doing is going to be the same as > using the same shash_desc struct for both the update and final calls. I > think we should be either using or adding a helper to the crypto api for > this. We're digging too deep into cryptoapi private structs with the > current usage. I think I found the solution. Instead of doing memset() + memcpy() + crypto_shash_update() + crypto_shash_final(), I could call crypto_shash_digest() which would wrap all of them for me in both the crc32c and sha256 case. Let me have a look and throw it in some testing. > Otherwise, thanks for doing this, it looks great overall. Thanks :) -- Johannes Thumshirn SUSE Labs Filesystems jthumshirn@xxxxxxx +49 911 74053 689 SUSE LINUX GmbH, Maxfeldstr. 5, 90409 Nürnberg GF: Felix Imendörffer, Mary Higgins, Sri Rasiah HRB 21284 (AG Nürnberg) Key fingerprint = EC38 9CAB C2C4 F25D 8600 D0D0 0393 969D 2D76 0850
